As complicated as our relationship with computers has been over the last half century, at least there is one constant: Wherever you find a team that is a swarm of insects. For users and managers have decades for weapons to use against infestations. Some try to structured programming tools, environments also supposedly sterile insects to reproduce. Others rely on Alpha ambitious and beta testing programs. But conceptually more satisfactory is the idea of getting equipment to detect and report (and perhaps even fix) their own mistakes automatically.
For some errors, automated testing requires a very high level of skill, a line in the true artificial intelligence. Not all these pests are out of reach, however, and products dedicated to their extermination began appearing on the market in the '80s. In general the work programs of the managers of the catch or stop creating typical user sessions (sequences of keystrokes and mouse movements) in pieces of software. Developers can run these sessions through the program being debugged and examine the error output. In the 90 automated test sector was sufficiently developed for us to publish a study on the technology ( "Bug Busters" CIO, March 1993).
Overall we were unimpressed. The tools seemed expensive, clumsy and lost a lot of trouble. The programs also presented a steep learning curve. "In the short term, the organizations implementing such products should expect lengthy production schedules, increased demand for development and a decline in software quality," he wrote. "IS managers are doing well, if a set of tools worth ... after three years."
A manager at the time might have been skeptical that the network systems of any scale could survive its own mistakes plague. Yet perhaps illogical, the development of the Internet turned out to be a boon both for the war against insects in general and automated testing in particular. The code sharing among developers was made easier, cheaper and faster, drawing considerably the efficiency of hunting manual errors. (The high reliability of open source software is a prime example of the benefits.) Patches was easier to distribute. Networks allow the installation of "flight recorders? Sensors that fall within the applications and submit reports of dysfunctional sessions back to the seller. "This is a big problem, because it means the technical support does not have to try to repeat the mistakes on their end," said Oliver Cole, president of systems of organized crime, a system of available vendor tools in Fairfax, Virginia Registrars flight sessions can also generate high quality evidence base for automated software testing. (Attest Technologies Fremont, California, produces such tools.)
0 comments:
Post a Comment